Work in progress on Surface Moisture modelling in the MDV

Work plan

0. write up some stuff I need to know about MDV surface moisture and its behaviour in this RMD as input to the paper

1. explore soil moisture dataset

   • data cleaning necessary?
   • understand surface moisture and temperature information distributions, determine cut-off value (0°C? -x°C?)
   • spatial configuration of loggers, i.e. think about validation strategy (how to do the spatial CV)

2. data gathering for pre-study: find a cloud free day and get as many useful spatial predictor datasets as possible

3. describe and understand relations between surface moisture and predictors & temperature

   • surface moisture and elevation
   • surface moisture and temperature

4. build model for case study

5. run workflow for the whole temporal and spatial setting

To do’s:

• ask Pierre about salinity and pH-paper progress

Paper relevant info

Introduction

Research Question:

• For Method 1 and 2: How accuarate can surface moisture be modeled for the open soil areas within the Antarctic Dry Valleys?

• actually interesting question: Which factors does the spatio-temporal surface moisture configuration in the MDV depend upon?

What we know about surface moisture in the MDV

• very low overall
• longer wavelenghts can pick it up
• short duration of hydrological events
• …

Questions

• should I only use values above 0 degrees? Find out when water freezes in the MDV and how the relation to salinity would be

Data

Calibration and validation

AWS and iButton spatial distribution

Already available:

• iButton data

To download:

• AWS quality controlled and official data (until 2019 or 2020) non quality controlled but full record until now

Example: Explorer’s Cove

There is:

• Air Temperature
• Precipitation
• Solar Radiation
• Relative Humidity (that’s what iButtons measure)
• Soil Moisture
• Soil Temperature
• Wind Direction and Speed

Potential spatial predictors

Potential spatial predictors

Already available:

• DEM 8 / 30m * TWI * slope
• rock outcrop to use as a mask
  • use soil type map to find out where there is only rock and no soil
• LST 30m
• soil types
• pH model (Pierre: “pH is very much correlated to soil moisture, and more linear than EC (because it’s already a log scale!), but I could also share EC estimates if need be.”)

To acquire

• RS data:
  • SWIR
  • downscaled LST
  • radar
• EC (Pierre)?

Methods

Method schematic overview

Discussion